[sigma relief]

Ticica, you are absolutely correct in terms of load capacity of a thin layer of epoxy. Linear rails are not very stiff and the majority of the load will be distributed within a few inches of each truck, but the calculations will still show plenty of strength.

I gues I didn't clearly state my concerns with self leveling epoxy, when used as a datum for low stress joints (ie rails) it is strong enough, but accuracy is still in question. For high stress areas (column mounting area on a traditional mill) or for areas subjected to point loads and abrasion (tables) it may not be ideal.

As far as self leveling epoxy goes, I have never worked with the material so I am only speculating on its properties. I have seen .003 in refrence to large setup plates for race cars and spacecraft. I would love to see the data sheets for meniscus formation in smaller areas however. I am assuming there is an area near the edge where the majority of the uncertanty takes place. If this areas is less than a few inches it can be worked around, but if it is too big, your castings will lave lots of useless material around the edges.

The bottom line is build your machine so the sum of all component tolerances is below your desired part accuracy. A flatness tolerance of .003 does not mean that the machine will not be able to beat that tolerance on many parts, but over the full travel, it will add significant unknowns.

[ticica]

As far as self leveling epoxy goes, I have never worked with the material so I am only speculating on its properties. I have seen .003 in refrence to large setup plates for race cars and spacecraft. I would love to see the data sheets for meniscus formation in smaller areas however. I am assuming there is an area near the edge where the majority of the uncertanty takes place. If this areas is less than a few inches it can be worked around, but if it is too big, your castings will lave lots of useless material around the edges.

As far as the meniscus effect goes... if one allows the mold to be higher than the top area of the flat surface (ie. not building a standard surface plate), then the sides of the mold near the top can be epoxy 'pre painted' with a brush. When the EG mix is poured, the EG epoxy will be in contact with the cured epoxy on the mold walls, as opposed to the release agent which may have a different surface tension. Wouldn't this effectivelly get rid of the meniscus? Are the surface tensions of dry and wet epoxy the same?

This might even work with 'inserts'. Wet them with epoxy and let it cure. There should be no meniscus when they are inserted into EG mix later.

Ivan

[ckelloug]

Sigma John,

I had the impression that the .003 was the error in the entire floor surface in the air force academy experiment which had a floor tens of feet long and wide. I think ticica's estimate of .0003 per foot is probably close. I plan to do an experiment on this one of these days as I have a 3' x 5' surface plate. All I need now is time and motivation.

Regards All,
Cameron

[altaic]

Hi all,
It's been awhile without any progress from me. The vacuum fitting that was delayed turned out to be the wrong part, so the company is sending me the correct part which should be here in a day or two. I've discovered that certain vacuum fittings are quite expensive and difficult to procure (I needed a 1-1/8-20 UNEF-2 thread to NW-25 ISO flange adaptor). If only I had a CNC lathe...

Anyway, I finalized my basic mold design, so I figured I'd post a rendering and a section-cut of it. The green bit is a linear vibrator, the yellow is a mounting plate for the vibrator so it can be removed without breaking the vacuum bag seal, the blue is the mold top, the red is the mold bottom, the orange are side panels for the mold bottom to allow for part removal (the sides of the mold are not angled) and for cheaper mold construction, the thin gray sheets are release-treated breather/bleeder fabric, the white connectors are nylon hose barbs (nylon should keep the connector free from cured epoxy), and the thin translucent layer around the mold is the vacuum bag. The E/G mixture goes in the void between the mold bottom and top/breather.

Additionally, I have been compiling data and notes in a wiki of mine, which I have cleaned up and edited for mass consumption. Is there a suitable wiki I can post all of this information? It would be excellent if CNCZone hosted a wiki. If not, I may just stick it on wikipedia.

Finally, I'm trying to work out the final density of the mixture we've been discussing. Attached are two tables showing the densities of the components. I'm multiplying each specific gravity with the % by mass and adding the products. I think this should give me the specific gravity of the aggregate without epoxy (i.e. not accounting for voids), but if we assume there will be 12% epoxy (i.e. 12% voids) I think I can get a fairly accurate number. The specific gravity of the aggregate without epoxy I calculated to be 2.94 g/ml. With 12% epoxy (with 1.08 g/ml density) per ml of aggregate, I get 1.08 * 0.12 = 0.1296 g to fill the voids in 1 ml of the aggregate. This gives me 3.07 g/ml (or kg/l) as a final density. Does that sound right?

:cheers:

[lgalla]

Hi Cameron,
You are correct.The Navel Academy surface plate is .003 over a 16 by 14 foot area.Granite surface plates are rated over a 2 sqft area.The .003 is not that bad considering the length.
Larry

[ckelloug]

Hi Will,

Sorry to be a bit slow responding. Your scheme looks good. Is this is a mold to build beams or a mold to build test parts? So you're putting a vacuum bag around the outside? I've been avoiding expensive vacuum flanges and just using ordinary tubing connectors. My valves are food grade valves which I think are used for controlling the flow of soda fountain syrup and I can say that I've also been using nylon and polyethylene fittings though I consider them disposable since I'm using .25 inch stuff and they're between a few cents and a few dollars a piece.

I believe that you have also calculated the mass percentages right. With any good luck, the packing model on which I based the distribution I suggested happens to be dead on and we actually get the 88% hoped for or better. The model's author said that it's supposed to be good to plus or minus 3% in his book. The only other note is that the model's density predictions can be skewed if the part being cast is less than 5 times the average dimension of the largest aggregate.

So, are you hoping to do what Tony (bloefeld) suggested and compact the mixture and then infuse the epoxy under vacuum pressure? I'm really hoping that that technique works as he described since it would very much simplify making large numbers of the same part with consistent properties. I also have some concern that additives might be necessary to make this sufficiently cooperative.

I don't think anybody has set up a wiki yet. I generally post my notes and observations here on the thread since the world's only dozen or so interested people seem to be here

I'm not sure the density computation you are using is quite the right one. If you have 1cc of material measured volumetrically packed as tightly as practical, I am assuming 88% of it will be rocks and 12 percent of the total volume will be voids. The caveat is that the volume that you get by dividing mass by specific gravity is a volume of the solid component, not the total volume. I'm thinking that the mass/total volume is then .88*(2.94g/1.12cc) +.12*1.08g/cc which gives me 2.44g/cc, not 3.04g/cc.

My justification for this is that if you get 2.94 g of the solid, it will have 1cc of solid portion but it's actually 1.12cc of space (aggregate plus voids) since the 12% voids are already built into it. Thus I think you end up having to correct the volume to be on a total space basis before you proceed with the density computation.

The mass times specific gravity argument gives volume of solids present, not total volume which has to be measured externally to get measured packing density. If you assume that the packing density is 88% then the actual density of the mixture has to have a density of less than the mass of 100% solids/divided by the volume of 100% solids. Using this logic, you end up with the mass of 100% solids divided by the volume it actually occupies which I assume to be 12% more than the solid volume by doing mass * specific weight.

It wasn't until greybeard asked one of his innocent sounding questions last year that I realized that this whole thing was more complicated than it looked and I didn't understand it the first time. Anybody have any comments on this argument or think I/we have it right this time?

Regards,

Cameron

[sigma relief]

I am not trying to keep bringing up an old topic, but I am still concerned about using epoxy alone as a datum. .003 flattness over 20 feet is not equal to .00015/foot, there will be a menuscus along the edges and depending on edge conditions, it may not be consistent all the way around. In the center of a large plate, the forces on the epoxy are uniform and it will suffer from only small deviations with large spacings. Near the edges, this is not the case and some small but measurable area will be less accurate, perhaps even more than the .003 stated in the specs. An automotive setup pad is only required to be accurate under and near the car. The 2ft appron is for convenience, some will be used for measurements, but the last few inches would most likely never be used.

I am actually excited about the prospects of using self leveling epoxy, but even in a large CNC machine, the unknown perimeter may be a sizable percentage of work space. At a minimum, it will require larger components and additional support castings that don't add value to the project.

Like anyone else googling self leveling epoxu, I have yet to find true scientific datasheets, only annecdotal tolerances of finished projects. If there is a "real" datasheet on any of the varieties of this stuff, I would love to see it.

John

[ckelloug]

sigma John,

You could very well be right: I don't think anybody has actually tested it here yet to have real data. I think everybody's belief in this including my own is based on thinking how nice it would be if it really worked. I seem to remember seeing a web page about a company that poured a floor for an aerospace application in europe but the details weren't given.

Conceptually, you should be able to get equivalent flatness to the curvature of the earth if the epoxy is thoroughly degassed and you are either far enough from the edge and/or the surface energies of the final layer and the initial layers are closely enough matched and nothing goes wrong.

I have already noticed a problem using leftover epoxy from epoxy formulation experiments in my lab: a tiny touch of amine blush on the surface of the epoxy can form a scum that makes the cast surface imperfect.

On the bright side, the epoxy I cast in a precision ground mold conformed to the mold surface so well that one could see sub .0001 variations in the original surface done by a commercial grinding shop and then hand lapped by me to 3000 grit diamond by the different ways light was reflected from the surface.

My current conclusion is that the replicant property of epoxy is very very good but the question of whether you can cast reliable datums is still open. I believe that there is some set of circumstances for which it will work better than a lot of tricks used by folks here building routers.

The question is whether you can use epoxy to produce either a surface flat for a mold component or actual machine ways for a metal working machine as good as hand scraped and lapped cast iron for lengths where the equipotential surface of the earth can be regarded as flat enough.

[lgalla]

Sigma John,
The 2ft appron you refer to is how a granite surface is rated.A.00015 granite plate is.00015 over 2 sq ft area.As the plate gets larger the tolerence is greater.
Regarding annecdotal tolerences and true scientific data sheets,The Navel Academy finds the epoxy plate good enough to replace their granite surface plate to duplicate zero gravity with air bearings.Is it not enough to beleive their findings?If .003 over 14X16ft is not good enough,then you may consider scraping cast iron.
Larry

[altaic]

Hey guys,

Cameron, that mold is just for test parts so I can try out different hardeners, pigments, etc. and measure dimensional distortion and possibly send you samples to test strength. The vacuum bag is just there so I don't have to worry about making my mold with complicated seals and such

The vacuum fittings caught me by surprise; I needed to somehow adapt 1 1/8-20 thread to 3/8" tubing. The only fitting I could find that fit the thread of my vacuum pump intake port was the ISO/KF/NW flange one. Had I a lathe of some sort I could have made a tube barb that threads right in for a few bucks. Oh well.

Re. mixture density-- I understood that the specific gravity is the density of the materials, not the packing density, IOW the voids are not considered. I did mess up the calculation, though. You're absolutely right that I had calculated 1.12 cc of mixture, instead of 1 cc, including .12 cc of epoxy. This is doubly wrong because the volume was wrong, and calculated epoxy content was .12/1.12 = 10.7%. However you did a similar sort of thing with your calculation, in that you took 88% of 2.94/1.12 and then added 12% of 1.08; I don't think the division by 1.12 cc should be there. I'm fairly confident it should just be 88% of 2.94 g/cc plus 12% of 1.08 g/cc, which is 2.72 g/cc.

Regarding the epoxy surface plate, I think the meniscus can be factored out by making the plate large and defining a usable area that is smaller. I'm working on making a mold for an E/G table top (that will also be a surface plate), and I'm planning on pouring a thin section of epoxy in a cavity that is a few inches wider/longer than my mold. Then I'll secure the mold in the center of the epoxy plate such that the epoxy plate forms the mold surface for the table top surface. I'm just guessing that a few inches will be enough for the meniscus to be negligible. The meniscus appears to taper off pretty rapidly with the epoxy I'm using.

Larry, the zero-gravity emulation with air bearings may not be the best example, since that necessitates the plate have a curve that matches gravity (or the imagined curve of the earth in that area) rather than be perfectly flat. In that case, self leveling epoxy is the ideal solution.

[the4thseal]

as far as the meniscus, we hammered that 200 or so posts ago but it an interesting topic. if you have read them and still concerened it is understandable. it is a very long thread.

[altaic]

I believe that you have also calculated the mass percentages right. With any good luck, the packing model on which I based the distribution I suggested happens to be dead on and we actually get the 88% hoped for or better. The model's author said that it's supposed to be good to plus or minus 3% in his book. The only other note is that the model's density predictions can be skewed if the part being cast is less than 5 times the average dimension of the largest aggregate.

So, are you hoping to do what Tony (bloefeld) suggested and compact the mixture and then infuse the epoxy under vacuum pressure? I'm really hoping that that technique works as he described since it would very much simplify making large numbers of the same part with consistent properties. I also have some concern that additives might be necessary to make this sufficiently cooperative.

The test mold isn't set up for infusion, although I think it could be done without too much trouble. In my mind, the part that makes E/G hard to reproduce is the fact that random packing is inconsistent. This is especially the case since the particles aren't spherical, and the larger ones aren't even cubic. How inconsistent, I have no idea, but you mentioned +/- 3%. That could be insignificant if the mold is set up to allow for one inaccurate dimension, or better yet, allow for a small cut-off section that can accommodate overflow (using 3% more mixture, for 0% to +6% error), a la stems/spouts used in metal casting. OTOH, it could be very bad with a mold not designed with this consideration, say one that is for a very long, thin cylinder being packed lengthwise; the variation in the volume will shorten or lengthen the cylinder greatly.

I may do some infusion experiments, though. In that case I'd have a resin inlet at the bottom and cover the inside bottom and top of the mold with breather fabric (unfortunately this would mean that two surfaces rather than one wouldn't be able to reproduce the mold's surface texture). Then I'd pour mixed epoxy in a pressure pot with a siphon tube that connects to the resin inlet on the mold, and pressurize the pot as much as possible. That way it's vacuumed into the mold as well as injected. Then when the proper amount has been injected, purge the vacuum and pressurize the mold as much as possible, forcing the epoxy into the vacuumed voids. It complicates the process a lot, and I probably wouldn't use infusion/injection for anything but mass production.

[altaic]

So, I'm starting to get really annoyed with Welch Vacuum / Gardener Denver. They still haven't sent me that part. I think they only have one tech who deals with RMAs and such, and he doesn't seem to be on top of things. Well, evidently he's not at all on top of things, since it's been two weeks and still no part; a part that he said would ship the next day. He even said he'd email me the RMA info and tracking number, but I got nothing. Eh, enough ranting.

I'm curious if you guys have thoughts on curing under pressure. Consider for a moment my mold set-up, where one entire surface is used to clamp via vacuum, and suppose this is pretty effective for evenly compacting the mixture. Now suppose instead of using a large clamping surface, we use a long rigid tube containing some mixture, with a cylinder atop it, i.e. a hydraulic multiplier. Vacuum pressure is only 14.7 psi; if we could degas (the tube included) using vacuum, and push the cylinder down by some means, we should easily be able to get an order of magnitude or more pressure. Of course, the mold would have to be able to withstand the pressure without deforming, but that aside: if the mix is being pressed with 150 psi (or even 500 or more psi), would there be elastic deformation in the particles that would, after the mix cures, cause internal stresses when the pressure is relaxed? OTOH, would the mixture be so much more dense that it would be stronger despite the stresses (if there are stresses)?

I'm trying to understand what happens as epoxy cures, and if pressure or vibrations are beneficial or detrimental... Or is it neither and such things alter the physical properties of the cured part to be better in some way and worse in another? For instance if a curing part were to be ultrasonically vibrated, would the epoxy form less rigid bonds and therefore allow for more elasticity and toughness in the macroscopic material?

[greybeard]

........Now suppose instead of using a large clamping surface, we use a long rigid tube containing some mixture, with a cylinder atop it, i.e. a hydraulic multiplier. Vacuum pressure is only 14.7 psi; if we could degas (the tube included) using vacuum, and push the cylinder down by some means, we should easily be able to get an order of magnitude or more pressure. Of course, the mold would have to be able to withstand the pressure without deforming, but that aside: if the mix is being pressed with 150 psi (or even 500 or more psi), would there be elastic deformation in the particles that would, after the mix cures, cause internal stresses when the pressure is relaxed? OTOH, would the mixture be so much more dense that it would be stronger despite the stresses (if there are stresses)?..........

Following your musings with interest, Altaic, as I see you might be approaching the effects of my spinning method which produced 200G at the periphery of the tube.
Unfortunately my mold distorted enough to make the thing a pain to correct each part afterwards, so it wnet on to the back burner.
Come the warmer weather I plan to give it another shot along the same lines, possibly under vac, so I'll be even closer to your idea.
Regards
John

[tootalew]

I started reading this post when it was started over two years ago by walter and some where lost interest, and even stopped visiting the zone for a long time. I never did lose interest in building a small benchtop cnc lathe, and have returned. I have read from the start this week the first 140 ish pages of this post and before going any further I just have to ask the question that Walter asked in post #1...has anyone built anything from epoxy granite yet? Ok so I changed the materail, but that was covered over two years ago.

[tootalew]

I do want to tip my hat to all those that have given so much time and effort to this thread on epoxy granite. It is just amazing reading. My lathe project is fairly simple so I am just going to use countertop grainte to start with. Sure would be nice to create a headstock or spindle housing using epoxy granite though. I will continue to read the rest of this thread, and maybe at sometime contribute. My thought is to just sink or attach real granite on top of big ole block of epoxy granite.

[probinson]

I'm trying to understand what happens as epoxy cures, and if pressure or vibrations are beneficial or detrimental... Or is it neither and such things alter the physical properties of the cured part to be better in some way and worse in another? For instance if a curing part were to be ultrasonically vibrated, would the epoxy form less rigid bonds and therefore allow for more elasticity and toughness in the macroscopic material?

I have not performed any testing to support this theory but, given that epoxy forms many more links and is much stronger when cured at elevated temperatures, I expect that low displacement ultrasonic vibration would likewise encourage increased link formation/entanglement and therefore also increase the bond strength.

Pete

[altaic]

I have not performed any testing to support this theory but, given that epoxy forms many more links and is much stronger when cured at elevated temperatures, I expect that low displacement ultrasonic vibration would likewise encourage increased link formation/entanglement and therefore also increase the bond strength.

Pete

Ah, that makes sense. I while ago I had read (and forgotten) that temperature affects cross linking and therefore strength and such. I was thinking that at some point physical manipulation during cure could make the bonds more elastic. Thinking about it more, though, it may just make the bonds tend toward the vector of the vibration rather than keep the cross-linked structure with lengthened chains. I now suspect elasticity would be better achieved using very small elastic particles dispersed in the epoxy.

One thing I've been ruminating on, and now believe I have a good understanding of, is how rigidity and vibration damping relate to material toughness. If you hit a block of material that is a (near) perfect vibration damper with a hammer, (almost) all of the energy of the hammer blow must be absorbed by the material. But being (almost) perfectly rigid, the structure of the material must not change. Thus, the energy must be absorbed as small molecular vibrations and heat, and there is only so much stress of this sort the molecules can take before breaking down. This is juxtaposed to a rigid non-damping material such as steel, where the hammer blow causes massive transmitted vibrations, some emanated as a loud ringing sound. I think the key is to design the material to dissipate the energy evenly throughout the block, which I believe is what E/G does.

Ideally, however, the material would have two states: one to absorb this energy up to a critical point, and the second to transmit the energy once past that point, before the material is over-loaded. IOW, damp up to a point, then transmit any additional energy. For instance, non-newtonian fluids (such as cornstarch and water) exhibit this type of behavior: its "natural" low-energy state is liquid, capable of absorbing/damping, and when hit with a hammer, the high energy area turns solid, transmitting energy. However, liquid pockets would weaken E/G. If there was instead a semi-elastic solid that hardens when stressed, though, that might be an interesting filler particle for our composite.

That reminds me about an article I read about epoxy fatigue... The gist of it was that epoxy develops small cracks/fractures as it fatigues, like most materials. However up to a point, these fractures are temporary and will "heal" when the stress is removed. Past that point the fatigue will become permanent. This gives epoxy fantastic fatigue resistance when stress magnitude and duration is limited and accounted for. IIRC, this article pointed out that Aluminum is the opposite; it has a determinable fatigue life, and any stress, small or large fatigues the metal permanently and will eventually lead to failure. More food for thought (and hopefully not rehashing something I missed in the thread)...

Edit: I just realized that such non-newtonian particles would likely cause structural change due to the transition from semi-solid to solid/crystalline or otherwise cause temperature instability. Since stability is an important part of E/G, that probably wouldn't be an acceptable tradeoff. Perhaps something like a shape memory alloy would allow temporary elasticity. Ferromagnetic shape memory alloys may be able to absorb mechanical energy and emit it as electro-magnetic waves. Of course, then the composite wouldn't have magnetic stability, but that might not be such a problem since E/G parts are unlikely to be in the presence of strong magnetic fields.

[ckelloug]

Will and Pete,

Ultrasonics are interesting but I've talked with Sonics and Materials and a suitable ultrasonic generator is in the $4000 range. They also pointed me at a paper in Macromolecular Materials and Engineering 2005, vol 290 pg 423-429. It turns out that there is a sonochemical effect applying ultrasound to epoxy.

Ultrasonic cavitation is powerful enough to cause ring opening reactions in epeoxy and other chemical damage to the side chains in the epoxy molecules. Up to a certain point however, this molecular damage increases the modulus and crosslink rate in the epoxy improving the cured properties. Too much ultrasound treatment however changes the chemistry enough to degrade the properties of the epoxy. Apparently, sonochemical effects also effect the hardener component and are much more likely to be detrimental. The research did not study ultrasonically treating the mixed epoxy as nobody wants to insert a $1700 dynamically tuned solid titatnium ultrasound probe into a material that will harden on it and perhaps detune or otherwise destroy it. The article found in some cases that the ultrasound treatment used to disperse carbon nanotubes into epoxy had as large an effect as the nanotubes themselves.

As to Will's earlier comment about pressure: The 1975 Proceedings of Polymers in Concrete paper by B.W. Staynes of Brighton Polytechnic suggests that setting under pressure improves the performance of epoxy concrete.

As far as improving the bonding, silane additives such as Dow Corning Z6040 and titanate additives such as Kenrich Petrochemicals LICA 34-J dramatically improve the bonding between epoxy and aggregate while simultaneously lowering mixture viscosity (sometimes dramatically). The viscosity effect comes from the blocking of hydrogen bonding sites on the aggregate surface preventing hydrogen bonds from altering the mixture rheology. I found some late 1970's handbooks of plastics fillers which explain some of this elegantly but I have not had time to fully read the material.

As for Will's comment about the vibration Damping, it will vary according to the glass transition temperature of the epoxy and the temperature at which the cast part is used. The glass transition is a function of both the chemistry of the epoxy and the curing temperature.

I spotted a 1950's book with a qualitative graph on vibration damping vs. glass transition temperature. (It's the book referenced as the authoritative reference on damping mechanisms in Slocum's <U>Precision Machine Design</U>). I don't remember the exact contents of the book but I seem to recall that there is a temperature region just below the glass transition temperature of the epoxy in which damping is maximized. Unfortunately, this is where stiffness starts to decay. On the upside, the reactive dilutants used to lower viscosity also lower the glass transitition temperature.

If we had some better data we could optimize the damping vs. the stiffness. I've been working to optimize particle content and flexural/compressive modulus and strength. Since the moduli of the epoxy and the particles are so different, there are wave reflections at every boundary which is one of the reasons that the damping is so good.

In response to Will's comment about non-newtonian fluids, cured epoxy is a non-newtonian fluid. It's behavior is highly dependent on strain rate. It also has a relatively abrupt phase change at the glass transition temperature where it becomes like rubber. Very heavily crosslinked epoxy behaves somewhat like a glass but slightly less crosslinked epoxy has some very very odd behavior. Based on a couple of undersized samples I performed D790 tests on, epoxy has a higher modulus at higher strain rates. This classifies it as a dilatant or shear thickening fluid: the same class as ooblek, (what I've heard the cornstarch and water fluid called), which Will talked about in his post.

In short, you guys (Will and Pete and John) are doing some first rate thinking. Now that I'm back from travel (at least some of it was travel to places with better tech libraries), I intend to take some packing density measurements on aggregates and get on with actually making some E/G and measuring the properties.

Good luck Will on getting the vacuum components assembled. Vaccum stuff is a royal pain in the posterior as I detailed in my sample mold posts a while back. I've had excellent service from Leybold Vacuum and I know that they have metric sized vacuum flanges if Welch is on your nerves.

Keep up the great work everybody. It's awesome to see all of this coming together.

Regards all,

Cameron

[Gizmot]

Will and Pete,

Ultrasonics are interesting but I've talked with Sonics and Materials and a suitable ultrasonic generator is in the $4000 range. They also pointed me at a paper in Macromolecular Materials and Engineering 2005, vol 290 pg 423-429. It turns out that there is a sonochemical effect applying ultrasound to epoxy.

Ultrasonic cavitation is powerful enough to cause ring opening reactions in epeoxy and other chemical damage to the side chains in the epoxy molecules. Up to a certain point however, this molecular damage increases the modulus and crosslink rate in the epoxy improving the cured properties. Too much ultrasound treatment however changes the chemistry enough to degrade the properties of the epoxy. Apparently, sonochemical effects also effect the hardener component and are much more likely to be detrimental. The research did not study ultrasonically treating the mixed epoxy as nobody wants to insert a $1700 dynamically tuned solid titatnium ultrasound probe into a material that will harden on it and perhaps detune or otherwise destroy it. The article found in some cases that the ultrasound treatment used to disperse carbon nanotubes into epoxy had as large an effect as the nanotubes themselves.

As to Will's earlier comment about pressure: The 1975 Proceedings of Polymers in Concrete paper by B.W. Staynes of Brighton Polytechnic suggests that setting under pressure improves the performance of epoxy concrete.

As far as improving the bonding, silane additives such as Dow Corning Z6040 and titanate additives such as Kenrich Petrochemicals LICA 34-J dramatically improve the bonding between epoxy and aggregate while simultaneously lowering mixture viscosity (sometimes dramatically). The viscosity effect comes from the blocking of hydrogen bonding sites on the aggregate surface preventing hydrogen bonds from altering the mixture rheology. I found some late 1970's handbooks of plastics fillers which explain some of this elegantly but I have not had time to fully read the material.

As for Will's comment about the vibration Damping, it will vary according to the glass transition temperature of the epoxy and the temperature at which the cast part is used. The glass transition is a function of both the chemistry of the epoxy and the curing temperature.

I spotted a 1950's book with a qualitative graph on vibration damping vs. glass transition temperature. (It's the book referenced as the authoritative reference on damping mechanisms in Slocum's <U>Precision Machine Design</U>). I don't remember the exact contents of the book but I seem to recall that there is a temperature region just below the glass transition temperature of the epoxy in which damping is maximized. Unfortunately, this is where stiffness starts to decay. On the upside, the reactive dilutants used to lower viscosity also lower the glass transitition temperature.

If we had some better data we could optimize the damping vs. the stiffness. I've been working to optimize particle content and flexural/compressive modulus and strength. Since the moduli of the epoxy and the particles are so different, there are wave reflections at every boundary which is one of the reasons that the damping is so good.

In response to Will's comment about non-newtonian fluids, cured epoxy is a non-newtonian fluid. It's behavior is highly dependent on strain rate. It also has a relatively abrupt phase change at the glass transition temperature where it becomes like rubber. Very heavily crosslinked epoxy behaves somewhat like a glass but slightly less crosslinked epoxy has some very very odd behavior. Based on a couple of undersized samples I performed D790 tests on, epoxy has a higher modulus at higher strain rates. This classifies it as a dilatant or shear thickening fluid: the same class as ooblek, (what I've heard the cornstarch and water fluid called), which Will talked about in his post.

In short, you guys (Will and Pete and John) are doing some first rate thinking. Now that I'm back from travel (at least some of it was travel to places with better tech libraries), I intend to take some packing density measurements on aggregates and get on with actually making some E/G and measuring the properties.

Good luck Will on getting the vacuum components assembled. Vaccum stuff is a royal pain in the posterior as I detailed in my sample mold posts a while back. I've had excellent service from Leybold Vacuum and I know that they have metric sized vacuum flanges if Welch is on your nerves.

Keep up the great work everybody. It's awesome to see all of this coming together.

Regards all,

Cameron

hey guys, I'm back from the grave.
Instead of wasting my own time to do research on the topic I'm now paid to do it

Anyway, I now work as a metrology engineer and I'm trying to acquire a 4x10 inspection surface plate. I just got a quote for a Starret granite and total is close to 20k, which is a good enough excuse to look at other solutions.
I'm trying to find out :
-what is epoxy wear resistance.
-what is the cost of creating a granite epoxy base
-what is the typical precision of countertop granite.

I'm open to all solution, including casting a large base out of concrete and having a precision floating top, buying a countertop granite plate.